There's some perceived wisdom where I work that when flowing $\ce{CO2}$ gas in pipes it acts somewhat "sticky", in the sense that it lingers around longer and takes more $\ce{N2}$ or air to purge out than, say, methane or $\ce{N2}$.

Unfortunately, no one here seems to know the physical or chemical basis for this observed result. I do have some quantitative data through various flow sensors and IR gas sensors that show it really does stay around longer than $\ce{N2}$ or $\ce{CH4}$, so I don't think it's due to bad observation.

As you can imagine, googling "why is $\ce{CO2}$ sticky" is not very helpful. If anyone can explain to me why this happens, I'd be very grateful, or even just point me in the right direction to do my own research.


Thanks for your responses so far.

Here's some more info on our gas system: we buy our $\ce{CO2}$ in large bottles at ~400 psi. We then feed it from outside through a pressure regulator that regulates it down to 2 bar (~29 psi) absolute pressure. All the pipes until this point are stainless steel.

It's then fed through a proportional valve and the pressure is held at 1030 mBar over a set of flow and gas sensors, then exhausted to atmosphere. The pipes in this section are a mix of Polyurethane tubes, and two proprietary rubber-like materials called Tygon and Viton tubing. These are held together with nylon fittings. All of this is at ambient temperature. In these pipes, only $\ce{N2}$, $\ce{CO2}$, and $\ce{CH4}$ flow.

Here are some graphs I collected of the system:

gas system test graph

In this graph, the green & brown lines are each gas being flowed in the system. The two blueish lines are a rough analog of how much $\ce{CO2}$ is detected in the system. As you can see, even after 5 mins of methane and N2 purging, and continuous intermittent $\ce{N2}$ purges, $\ce{CO2}$ still creeps back into the system when no gas is flowing.

Here's a Graph of Gas Flow

gas flow graph

The [Flow In] sensor is just after the 1030 Prop valve, and the [Flow Out] sensor is just before the exhaust. They are both identical Wheatstone bridge thermal flow sensors.

As you can see, there's a marked increase in flow loss (Flow In/Flow Out) from 2% to about 7% when CO2 is flowing, and some very strange transient spikes when I switch gasses to/from $\ce{CO2}$.

That's most of the data I've collected. Anecdotally, I've also noticed that the $\ce{CO2}$ valve is significantly harder to turn than the $\ce{N2}$ or $\ce{CH4}$ valves.

I hope this helps with the problem, and thank you again for your help so far.

  • $\begingroup$ There are some cross-country CO2 pipelines ( like 18" diameter). They pose unusual fracture concerns for the steel pipe. While following the technology , I never encountered "sticky" considerations. Operating pressures are above the critical point. $\endgroup$ Feb 16 at 23:13
  • $\begingroup$ @Andrew I have added some more supporting information. Let me know what you think. $\endgroup$ Feb 17 at 10:12

1 Answer 1


$\ce{CO2}$ solubility in water is about 100 times higher than other gasses. So if there is any moisture on the walls, $\ce{CO2}$ is much more likely to stay there.

$\ce{CO2}$ is also much more chemically active, forming lots of compounds through its carbonic acid. Some of these compounds are reversible, can absorb and emit $\ce{CO2}$ as its partial pressure or temperature changes.

There is likely water or mineral contamination in your system.

  • 11
    $\begingroup$ Indeed, unless you can pump out the lines to less than 1E-6 Torr, you have water adsorbed on the walls of your piping. $\endgroup$
    – Jon Custer
    Feb 16 at 18:00
  • 3
    $\begingroup$ @JonCuster baking the lines helps too, but I suspect the OP's system isn't set up for that and it would be harder than getting a turbopump on there. The materials may not be suitable for proper heating anyway. Even with a pump it would be much quicker with some heat. $\endgroup$
    – Chris H
    Feb 17 at 12:00
  • 1
    $\begingroup$ Also with a vacuum pump and voids separated by long runs, a process of pump out, N2 purge, pump out etc. can be more effective than simple pumping (He purge in my case as N2 is one thing I'm trying to get rid of to avoid it freezing). And optimising changeover processes so ambient air never gets in $\endgroup$
    – Chris H
    Feb 17 at 12:02
  • 3
    $\begingroup$ @ChrisH - Pump/purge cycles work wonders too... So many tricks of the trade out there, and so few places to learn them anymore... I just queried an editor at the Journal of Vacuum Science and Technology if they would be interested in a vacuum technique article based on some work we did on a project. They did say OK, but that there was no identified area editor for those anymore... $\endgroup$
    – Jon Custer
    Feb 17 at 14:23
  • 1
    $\begingroup$ @JonCuster a varied engineering/science career has left me well placed to learn on the job - just as well as there's no other way. But we need the time and space to fiddle about without causing big problems and that opportunity seems hard to find these days $\endgroup$
    – Chris H
    Feb 17 at 15:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.